removing false positives is the biggest challenge, we use layers of filtering using JUDGE&JURY system still the findings that pass the filter need to have a passing POC for 100% validation. Best part? all of it is opensource for the community just bring your API keys ><
https://t.co/96D78W17cy
what if instead of one LLM reviewing a smart contract vulnerability
you had 4 LLMs actively arguing about it
skeptic vs attacker vs auditor vs judge
that's literally what i built
@delveroin hi, Critikal is a fully autonomous web3 security research agent. it can autonomously detect and exploit bugs in web3 protocols and the best part? its fully open source
https://t.co/sxF6ItmA11
@pashov https://t.co/sxF6ItmA11
we have been working on this open source Autonomous AI agent which secures web3, give it a repo link and itll help you in your security research
Critikal is now open source.
An autonomous AI agent that hacks smart contracts.
Hand it a repo URL → it recons the protocol, maps the attack surface, forms exploit hypotheses, and writes Foundry PoCs that actually pass.
MIT. Built solo.
https://t.co/NMwnI02LfB